Many types of laser scanners and printers use a one-dimensional (1-D) beam deflector such as a spinning polygon, monogon, hologon, or a reciprocating galvo mirror, that scans the beam along a line. To provide two-dimensional (2-D) scanning with a 1-D beam deflector requires that mechanical motion be provided for scanning in the other dimension, orthogonal to the first. This is commonly done by a transport system that moves the scanned media or moves the scanning deflector.
Two-dimensional beam deflectors provide a simpler solution for steering the light beam for scanning along two orthogonal directions or axes. With some scanning laser systems, this is often termed Two-Axis Beam Steering (TABS). TABS scanning can use an arrangement comprised of two galvo mirrors like the ones made by Cambridge Technology, Cambridge Mass.; GSI Lumonics (previously General Scanning Inc) Bedford Mass.; Nuffield Technology Inc., Windham N.H.; and GalvoScan LLC, South Royalton Vt. Similar Fast Scanning Mirrors (FSM) systems also steer the beam in two dimensions using reciprocating reflective surfaces. Galvo mirrors are commonly used as deflectors due to their relatively wide deflection angles and high scan speed, especially when used in the resonant mode. Because galvo scanners scan only in one direction, a pair of galvo mirrors in series is used as the deflection system to accomplish 2-D scanning.
Laser beam scanning systems are generally classified by the arrangement of the deflection system relative to the focusing lens. When deflection system components follow the focusing lens, the system is termed a post-objective system. When the focusing or scan lens follows the beam deflector, the system is called a pre-objective scanner. Post objective scanners are usually simpler in design compared with pre-objective scanners, but are generally more limited in terms of scan fields and are more prone to distortion and field curvature.
The schematic diagram of FIG. 1 shows how light is directed for 2-D scanning by components in a pre-objective scanning system 80. A light beam 82 of beam width B, preferably collimated, is deflected by a first galvo mirror 35 that scans relative to a first axis and toward a second galvo mirror 35A that scans relative to a second axis that is orthogonal to the first axis. A scan lens 120 then directs this 2-D scan to form a 2-D image 130.
Among disadvantages of the arrangement of FIG. 1, second mirror 35A must be large enough to accommodate the deflected beam from first mirror 35, and thus cannot operate at high speed. Another disadvantage relates to mirror positioning. With a pre-objective system, the beam deflector provides the best optical performance when it is positioned in an external entrance pupil of the scan lens. This is shown as pupil 30 in FIG. 1. However, for such a two-mirror system, this would require that both mirrors 35 and 35A be positioned at entrance pupil 30. As a compromise, galvo mirrors 35 and 35A are generally positioned close to, and equidistant from, pupil 30, displaced at a distance L as shown in FIG. 1. When this is done, because both mirrors 35 and 35A are displaced from entrance pupil 30, scan lens 120 must have an aperture large enough to accommodate the beam displacements. This adds cost and size, requiring that scan lens 120 have a higher effective numerical aperture than does a system that uses a single two-dimensional scan mirror.
With both mirrors displaced from the entrance pupil, the aperture diameter D of the aperture corresponding to entrance pupil 30 is given by:D=B+2*L*Tan (α)Where:B is the beam diameter of light beam 82;L is the distance from pupil 30, along the axis, of the farthest mirror galvo;and α is the semi beam angle.
For example, with a beam diameter of 10 mm, a shift L of 20 mm, and a semi scan angle α of 15 degrees, the aperture D is 20.7 mm. Thus, the entrance pupil must be about twice the diameter of the beam. The numerical aperture (NA) of the scan lens 120 is therefore twice the NA of a scan lens where the mirror deflector is located at the entrance pupil 30.
As is shown in FIG. 2, one approach to solve this problem and reduce the NA of scan lens 120 is to optically co-locate galvo scanning mirrors 35 and 35A. There can be a number of ways to do this using refractive and reflective relay optics. Referring to the schematic diagram of FIG. 2, there is shown an example of an optical relay 90 that relays galvo mirror 35 onto galvo mirror 35A. In the arrangement of FIG. 2, the pupil relay uses two off-axis concave mirrors 55 and 56 to relay galvo mirror 35 onto galvo mirror 35A. With this type of solution, both deflectors can thus be optically positioned within the entrance pupil of scanning lens 120. This reduces the numerical aperture requirements for lens 120, as described earlier, and reduces the size requirements of the second galvo mirror 35A. Actuators 32 and 32A control the rotation of their respective scanning mirrors 35 and 35A.
In spite of its advantages for reducing size and performance requirements of other components in the optical system, however, the arrangement of optical relay 90 as shown in FIG. 2 has a number of problems that prevent its use in most laser scanner applications. Off-axis aberrations of the concave mirrors can seriously degrade the performance of such a system. Both mirrors are relatively large, requiring precision manufacture to minimize defects in maintaining exact curvatures. This solution is not particularly compact and does not scale well for large scan angles.
As exhibited in the example of FIG. 2, pupil relay optics show some promise for at least reducing some of the problems inherent to 2-D beam scanners using lasers. However, problems remain. In order for a pupil relay to satisfactorily serve 2-D beam scanning applications, the following basic set of requirements must be met:                (i) Low aberration. While some amount of aberration is inevitable, it is important that the pupil relay solution be well corrected and have minimal aberration.        (ii) Capable of handling large deflection angles. Angles of up to 12 degrees and larger should be accommodated.        (iii) Preserves the phase of the beam wavefront. When this requirement is met, a collimated input beam with a planar phase wavefront that enters the entrance pupil of the relay, on axis or at an angle within its specified field of view, exits from the exit pupil as a collimated beam, again with planar phase wavefront. The optical path difference (OPD) between any point at the entrance pupil and its conjugate point at the exit pupil is the same. This characteristic is of particular interest for laser scanning. The capability to preserve the phase of the beam wavefront distinguishes the performance requirements of a pupil relay system from the requirements of an imaging relay system. In an imaging relay system, beam wavefront and phase considerations are unimportant and the phase of the beam wavefront is not preserved.        (iv) Capable of providing a large pupil size, effectively forming an image of a circular pupil.        (v) Afocal. For beam relay applications, it is most preferable to handle collimated light. Exit and entrance pupils should be at infinity.        (vi) Color-corrected. This requirement depends on the application. Good color correction enables use of either monochromatic light or polychromatic light over a broad spectral range.        (vii) Low cost. This relates both to precision of assembly and number of components.        (viii) Reduced size.        
However, as seen from the example of FIG. 2, pupil relay solutions proposed thus far fail to satisfy all of these requirements. Instead, conventional pupil relay solutions typically compromise on one or more of these basic requirements.
In a paper entitled “Offner-type pupil relay optics for a scanning system” by G. C. deWit and J. J. M. Braat, in Design and Engineering of Optical Systems, SPIE vol. 2774, pp. 553-561, three possible pupil relay solutions are compared, including a spherical mirror on-axis relay, a spherical mirror off-axis relay, and an Offner-type system. These researchers conclude that the spherical mirror on-axis solution is optically superior to the other two proposed solutions.
Significantly, researchers deWit and Braat were intrigued with some of the advantages of an Offner-type solution, but were unable to make this arrangement work satisfactorily as a pupil relay and found the Offner-type arrangement inferior to on-axis spherical mirror designs. The authors cite the advantages of Offner optics as they relate to size, optical properties, and unlimited horizontal field of view (FOV). However, the Offner arrangement does not provide a pupil relay and is, by itself, a poor solution for directing a 2-D scanning beam, chiefly because it fails to preserve the beam wavefront, a significant needed feature of a pupil relay as noted earlier in requirement (iii).
It is instructive to understand more clearly why the Offner optical system, disclosed in U.S. Pat. No. 3,748,015 entitled “Unit Power Imaging Catoptric Anastigmat” to Offner, does not function as a pupil relay. This shortcoming is most readily shown by a description of the Offner optical system itself. Referring to FIG. 3, an Offner optical system 92 is a one-to-one object-to-image relay system using two concentric mirrors, a primary concave mirror 50 and a secondary convex mirror 60. The system is afocal, with its entrance pupil at infinity. The aperture stop of this optical system, pupil 65, is at secondary convex mirror 60. This system is corrected for all third order aberrations and for a number of higher order aberrations.
The imaging function of Offner optics is constrained to image a specific field of a particular shape, rather than for a circular beam. The Offner optical system, as shown in the example of FIG. 3, is a member of a class of imaging systems that have an annular object, or a ring object 15 and, in turn, form a corresponding ring image 15A. Because of higher order aberrations, such as fifth order astigmatism, the object shape is limited to a thin arcuate region 15 about the optical axis OA. Thus, in practice, the Offner system is used to scan across a 2-D area and form an image of the arcuate object shown as object 15 in FIG. 3. Examples of how the Offner optics are used for scanning are given, for example in U.S. Pat. No. 5,221,975 entitled “High Resolution Scanner” to Kessler that describes a CCD scanner for film reproduction and in U.S. Pat. No. 6,304,315 entitled “High Speed High Resolution Continuous Optical Film Printer for Duplication Motion Films” to Kessler et al. As each of these patents shows, the object of the Offner optical system is an arc of limited thickness and the optical system faithfully images that arc with little aberration, effectively scanning a 2-D image over this arcuate image area.
In an attempt to improve the relative size of the arc that can be imaged by this two-mirror system, Suzuki in U.S. Pat. No. 4,097,125 adjusts the positions of the two curved mirrors so that they are no longer concentric, as is required in the conventional Offner system. Even with this change, however, the width of the relatively narrow slit imaged by the Offner system can only be increased by about a factor of 2. The modified arrangement provided by Suzuki '125 still images an arcuate object, not a circular beam, and would also fail as a pupil relay.
Another significant problem with Offner-type optics relates to the beam wavefront. When the Offner system is used as an object-to-image relay, the object points are incoherent with each other. This is not a concern for imaging, as was noted earlier. However, in order to maintain beam quality as a beam relay, the beam phase wavefront at the entrance pupil must be preserved at the exit pupil. This is not the case with Offner optics. When the Offner system is used as an object-to-image relay, five of the third-order aberrations, namely spherical, coma, astigmatism, Petzval, and distortion, are corrected. However, one of the third order aberrations, called “piston error” which particularly relates to the phase difference between different object points, is not corrected. In the type of imaging system for which the Offner optics are designed, there is no need to correct for the piston error since the object points are themselves generally incoherent with each other and phase is not important. However, for beam relay optics, such a phase difference is an aberration that severely degrades beam quality for a coherent beam and can render the optical system unusable. In summary, because it is an image relay system, and not a pupil relay system, as described in requirement (iii) above, the Offner system does not preserve phase relationships.
Thus, it can be appreciated by those skilled in the optical arts that the Offner system is an afocal imaging relay, not an afocal pupil relay, and it is no surprise that, in spite of their interest in some of the potential capabilities and compactness of Offner optics, researchers have been unable to adapt this arrangement for use as an acceptable pupil relay.
The need remains for pupil relay optics for 2-D scanning that meet requirements (i) to (viii) given earlier. In spite of some perceived advantages, however, solutions posed thus far have failed to take advantage of Offner-type optics for use as a beam relay for this purpose, due to inherent limitations of such systems.